]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
17 years agoAdd hypervisor interfaces for logical domain channels from the hypervisor API docs
Kip Macy [Thu, 23 Nov 2006 03:52:39 +0000 (03:52 +0000)]
Add hypervisor interfaces for logical domain channels from the hypervisor API docs
remove bogus CDDL

17 years agoIn contrast to the non-obvious and flexible nature of the optimized bcopy in t1_copy...
Kip Macy [Thu, 23 Nov 2006 02:25:16 +0000 (02:25 +0000)]
In contrast to the non-obvious and flexible nature of the optimized bcopy in t1_copy.S (which
shall retain its CDDL copyright, and thus likely be removed from GENERIC) I have removed the CDDL
from hcall.S because there is zero flexibility in the implementation of hypercalls as they derive
directly from the hypervisor interface which is not copyrighted (ironically the source for the
hypervisor itself is BSD licensed).

It is best to start any bikeshed about this as soon as possible.

Discussed with: bsdimp

17 years agoRemove mount_devfs, mount_ext2fs, mount_fdescfs, mount_linprocfs,
Craig Rodrigues [Thu, 23 Nov 2006 01:44:15 +0000 (01:44 +0000)]
Remove mount_devfs, mount_ext2fs, mount_fdescfs, mount_linprocfs,
mount_procfs, mount_std, which are now obsolete.

Requested by: netchild, ache

17 years agoAdd a vendor ID for O2Micro, obtained from
Doug Barton [Thu, 23 Nov 2006 00:58:40 +0000 (00:58 +0000)]
Add a vendor ID for O2Micro, obtained from
http://www.usb.org/developers/tools

Add a product ID for the Dell TrueMobile 350 Bluetooth USB Adapter
obtained from NetBSD's usbdevs file.

17 years agoOops - what people refer to as linux_base-fc4 calls itself linux_base-fc
Ken Smith [Thu, 23 Nov 2006 00:57:10 +0000 (00:57 +0000)]
Oops - what people refer to as linux_base-fc4 calls itself linux_base-fc
and that's what pkg_add needs.

MFC after: 1 day

17 years agoFix TSO support on sun4v
Kip Macy [Thu, 23 Nov 2006 00:52:52 +0000 (00:52 +0000)]
Fix TSO support on sun4v

- incorporate csjp's fix for a mishandled endian conversion
- convert PAGE_SIZE to 4096 for PCIe adapter workaround (my page size is not 4k)
- implement em_read_pcie_cap_reg where we set the max read size on pcie to 4k (taken from mxge)

Reviewed by: scottl and jfvogel

17 years agoadd codes for Atheros USB devices; shuffle one ural code to
Sam Leffler [Thu, 23 Nov 2006 00:50:48 +0000 (00:50 +0000)]
add codes for Atheros USB devices; shuffle one ural code to
avoid conflict

Obtained from: openbsd
MFC after: 1 month

17 years agoAdd USR997902 to the list of devices supported by re(4).
Pyun YongHyeon [Thu, 23 Nov 2006 00:28:58 +0000 (00:28 +0000)]
Add USR997902 to the list of devices supported by re(4).

17 years agoAdd option -x to display output values in hex in addition to decimal.
Greg Lehey [Thu, 23 Nov 2006 00:20:54 +0000 (00:20 +0000)]
Add option -x to display output values in hex in addition to decimal.

17 years agoFix a race in soclose() where connections could be queued to the
Mohan Srinivasan [Wed, 22 Nov 2006 23:54:29 +0000 (23:54 +0000)]
Fix a race in soclose() where connections could be queued to the
listening socket after the pass that cleans those queues. This
results in these connections being orphaned (and leaked). The fix
is to clean up the so queues after detaching the socket from the
protocol. Thanks to ups and jhb for discussions and a thorough code
review.

17 years agoMention that the following binaries have been detached
Craig Rodrigues [Wed, 22 Nov 2006 23:01:40 +0000 (23:01 +0000)]
Mention that the following binaries have been detached
from the build: mount_devfs, mount_ext2fs,
mount_fdescfs, mount_procfs, mount_linprocfs, and mount_std.

17 years agoDetach mount_devfs, mount_ext2fs, mount_fdescfs, mount_procfs,
Craig Rodrigues [Wed, 22 Nov 2006 22:58:10 +0000 (22:58 +0000)]
Detach mount_devfs, mount_ext2fs, mount_fdescfs, mount_procfs,
mount_linprocfs, and mount_std from the build.  They are no longer
used, and can be replaced with "mount -t fstype".

17 years agoRemove mount_ext2fs.
Craig Rodrigues [Wed, 22 Nov 2006 22:55:54 +0000 (22:55 +0000)]
Remove mount_ext2fs.

17 years agoUse our standard section 4 SYNOPSIS.
Christian Brueffer [Wed, 22 Nov 2006 21:30:02 +0000 (21:30 +0000)]
Use our standard section 4 SYNOPSIS.

MFC after: 3 days

17 years agoo Check that -w width is not above maximum. Use DWIDTH instead of
Maxim Konovalov [Wed, 22 Nov 2006 21:05:17 +0000 (21:05 +0000)]
o Check that -w width is not above maximum.  Use DWIDTH instead of
hardcoded constant.

Reported by: Gruzicki Wlodek on BugTraq
Obtained from: NetBSD, OpenBSD
MFC after: 1 week

17 years agoFix transposition of width and value arguments to pci_config_write()
Andrew Gallatin [Wed, 22 Nov 2006 18:47:24 +0000 (18:47 +0000)]
Fix transposition of width and value arguments to pci_config_write()
when setting up the read request size.

Pointed out by: kmacy

17 years agoFix coherency issue. From submitter:
Warner Losh [Wed, 22 Nov 2006 17:56:36 +0000 (17:56 +0000)]
Fix coherency issue.  From submitter:

I have been debugging the usb problems some more. Your were
right in your assumption (thanks for the pointer) about lack
of calls to bus_dmamap_sync().  In usbdi.c bus_dmamap_sync()
does get used for transfers that move data from PC to USB and
it is used for transfers that move data from USB to PC. But
someone forgot that control transfers consist of possibly two
data chunks : the request itself and optionally a buffer of
data that should be transfered to or from the USB device.  On
requests to the control endpoint without additional data
bus_dmamap_sync() didn't get called. For some reason my first
tests with umass worked (due to enough cache poisening I
guess).  The attached patch adds a call to bus_dmamap_sync()
to usbdi.c and now all devices I have tried work out of the
box.  I have successfully transfered large files using the
if_axe driver and I have mounted several different umass
devices.

submitted by: Daan Vreeken
sponsored by: Vitsch Electronics
reviewed by: cognet@

17 years agoChange error codes returned by protocol operations when an inpcb is
Sam Leffler [Wed, 22 Nov 2006 17:16:54 +0000 (17:16 +0000)]
Change error codes returned by protocol operations when an inpcb is
marked INP_DROPPED or INP_TIMEWAIT:
o return ECONNRESET instead of EINVAL for close, disconnect, shutdown,
  rcvd, rcvoob, and send operations
o return ECONNABORTED instead of EINVAL for accept

These changes should reduce confusion in applications since EINVAL is
normally interpreted to mean an invalid file descriptor.  This change
does not conflict with POSIX or other standards I checked. The return
of EINVAL has always been possible but rare; it's become more common
with recent changes to the socket/inpcb handling and with finer-grained
locking and preemption.

Note: there are other instances of EINVAL for this state that were
      left unchanged; they should be reviewed.

Reviewed by: rwatson, andre, ru
MFC after: 1 month

17 years agoInitialization bugfixes and enhancements:
Andrew Gallatin [Wed, 22 Nov 2006 16:33:40 +0000 (16:33 +0000)]
Initialization bugfixes and enhancements:

- Fix bug preventing adoption of running firmware
- Set PCIe max read request size to 4KB
- Read PCIe link width from config space
- Assume aligned completions from the southbridge ports
  of intel E5000 chips
- Use aligned firmware when link width is x4 or less
- Add hw.mxge.force_firmware tunable to allow user to force
  selection of aligned (or unaligned) firmware

17 years agoRemove reference to mount_fdescfs.
Craig Rodrigues [Wed, 22 Nov 2006 15:52:24 +0000 (15:52 +0000)]
Remove reference to mount_fdescfs.

Suggested by: ru

17 years agoCross-reference nextboot(8).
Joseph Koshy [Wed, 22 Nov 2006 13:12:34 +0000 (13:12 +0000)]
Cross-reference nextboot(8).

MFC after: 3 days

17 years agoBring in status led support for /dev/led/gpioled on Avila.
Kevin Lo [Wed, 22 Nov 2006 12:57:17 +0000 (12:57 +0000)]
Bring in status led support for /dev/led/gpioled on Avila.

Approved by: cognet

17 years agoRemove reference to mount_devfs(8).
Craig Rodrigues [Wed, 22 Nov 2006 11:45:06 +0000 (11:45 +0000)]
Remove reference to mount_devfs(8).

Reminded by: ru

17 years agoIt's been possible to build linprocfs as a module for some time now.
Ruslan Ermilov [Wed, 22 Nov 2006 10:34:12 +0000 (10:34 +0000)]
It's been possible to build linprocfs as a module for some time now.

Submitted by: rdivacky

17 years agoAdd the SUBMAKE_TARGETS (like ppp_make) to the dependency list
John Birrell [Wed, 22 Nov 2006 07:52:14 +0000 (07:52 +0000)]
Add the SUBMAKE_TARGETS (like ppp_make) to the dependency list
for linking the crunched app so that when a parallel make is run
(like -j32 on sun4v), the link waits for the sub-make processes
to complete.

17 years agoMFp4: Make it work :-)
Warner Losh [Wed, 22 Nov 2006 06:51:59 +0000 (06:51 +0000)]
MFp4: Make it work :-)
o Don't delay when checking the done bits.  There's no gain other
  than a small performance hit.
o calculate the clock divisors better (things are still way slow,
  so maybe there's more here?)
o don't always fail reset.  Always succeed instead.
o fix inverted logic around at91_twi_wait() return value
o remove debug code
o remove unneeded, unworking junk

17 years agoMFp4: Add ixpiic bit-bang driver.
Warner Losh [Wed, 22 Nov 2006 06:41:37 +0000 (06:41 +0000)]
MFp4: Add ixpiic bit-bang driver.

Submitted by: sam@

17 years ago<blush> copyout on read, not write.
Warner Losh [Wed, 22 Nov 2006 06:03:39 +0000 (06:03 +0000)]
<blush> copyout on read, not write.
Tweak a comment while I'm here.

17 years agoIntegrate, but do not enable support for dynamically resizing TSBs
Kip Macy [Wed, 22 Nov 2006 05:54:24 +0000 (05:54 +0000)]
Integrate, but do not enable support for dynamically resizing TSBs

17 years agopmap_track_modified has been removed from other architectures -
Kip Macy [Wed, 22 Nov 2006 04:50:55 +0000 (04:50 +0000)]
pmap_track_modified has been removed from other architectures -
likewise remove from sun4v

17 years agoreduce whining from LINT by removing another GPL sound driver
Kip Macy [Wed, 22 Nov 2006 04:35:58 +0000 (04:35 +0000)]
reduce whining from LINT by removing another GPL sound driver

17 years agoadd support for resizing the the tte_hash of multi-threaded processes
Kip Macy [Wed, 22 Nov 2006 04:33:34 +0000 (04:33 +0000)]
add support for resizing the the tte_hash of multi-threaded processes

17 years agoremove unused field from pcpu structure
Kip Macy [Wed, 22 Nov 2006 04:27:24 +0000 (04:27 +0000)]
remove unused field from pcpu structure

17 years agoremove dead code from tsb.c
Kip Macy [Wed, 22 Nov 2006 04:13:30 +0000 (04:13 +0000)]
remove dead code from tsb.c
switch tsbscratch over to using order of number of pages as opposed to actual number of pages
switch tsb.c over to using wrappers for contig page allocation

17 years agomove contiguous allocation and free routines from tte_hash.c into pmap.c
Kip Macy [Wed, 22 Nov 2006 03:35:37 +0000 (03:35 +0000)]
move contiguous allocation and free routines from tte_hash.c into pmap.c

17 years agoAdd tte_hash and tsb update handlers for handling tte_hash and tsb
Kip Macy [Wed, 22 Nov 2006 01:47:58 +0000 (01:47 +0000)]
Add tte_hash and tsb update handlers for handling tte_hash and tsb
resizing across cpus

17 years agoRemove reference to mount_fdescfs.
Craig Rodrigues [Wed, 22 Nov 2006 00:34:17 +0000 (00:34 +0000)]
Remove reference to mount_fdescfs.

Reminded by: ru

17 years agoRemove references to mount_std, mount_procfs.
Craig Rodrigues [Wed, 22 Nov 2006 00:33:19 +0000 (00:33 +0000)]
Remove references to mount_std, mount_procfs.

Reminded by: ru

17 years agoRemove references to mount_devfs, mount_fdescfs, mount_linprocfs,
Craig Rodrigues [Wed, 22 Nov 2006 00:26:55 +0000 (00:26 +0000)]
Remove references to mount_devfs, mount_fdescfs, mount_linprocfs,
mount_procfs, and mount_std.

Reminded by: ru

17 years agoRemove references to mount_devfs, mount_fdescfs, mount_linprocfs, mount_procfs.
Craig Rodrigues [Wed, 22 Nov 2006 00:23:11 +0000 (00:23 +0000)]
Remove references to mount_devfs, mount_fdescfs, mount_linprocfs, mount_procfs.

Reminded by: ru

17 years agoAdd mechanism to track TSB misses in tsb miss handler
Kip Macy [Wed, 22 Nov 2006 00:18:22 +0000 (00:18 +0000)]
Add mechanism to track TSB misses in tsb miss handler
Remove unused debug code

17 years agoRemove reference to mount_procfs.
Craig Rodrigues [Wed, 22 Nov 2006 00:08:35 +0000 (00:08 +0000)]
Remove reference to mount_procfs.

Reminded by: ru

17 years agoChange mount_devfs reference to "mount -t devfs".
Craig Rodrigues [Tue, 21 Nov 2006 23:45:44 +0000 (23:45 +0000)]
Change mount_devfs reference to "mount -t devfs".

Reminded by: ru

17 years agoAcknowledge (dearly) departed files
Doug Barton [Tue, 21 Nov 2006 22:29:06 +0000 (22:29 +0000)]
Acknowledge (dearly) departed files

17 years agoAdded MSI support.
Andrew Gallatin [Tue, 21 Nov 2006 15:55:01 +0000 (15:55 +0000)]
Added MSI support.

Sponsored by: Myricom Inc.

17 years agoRemove the NO_CTF option. It would have been required to support
John Birrell [Tue, 21 Nov 2006 08:25:28 +0000 (08:25 +0000)]
Remove the NO_CTF option. It would have been required to support
DTrace.

17 years agoRemove the KDTRACE option because I can't implement it the
John Birrell [Tue, 21 Nov 2006 08:23:20 +0000 (08:23 +0000)]
Remove the KDTRACE option because I can't implement it the
way I intended due to licensing restrictions. I had intended
that it would be defaulted on, with opt-out possible for
companies that don't accept the CDDL. The FreeBSD GENERIC
kernel has to be entirely BSD licensed, so the only alternative
would have been to make KDTRACE an opt-in option. That isn't
a design I favour.

17 years agoo Sort Xrs.
Maxim Konovalov [Tue, 21 Nov 2006 06:52:55 +0000 (06:52 +0000)]
o Sort Xrs.

17 years agodon't mark the cksum as invalid here... off is incorrect when we get
John-Mark Gurney [Tue, 21 Nov 2006 05:46:09 +0000 (05:46 +0000)]
don't mark the cksum as invalid here...  off is incorrect when we get
here, it's either unset, or it's valid, so we don't need to do anything
different...

Reported by: Neterion (via rwatson)

17 years agoChange reference from mount(2) to nmount(2), since mount(8) uses
Craig Rodrigues [Tue, 21 Nov 2006 05:45:14 +0000 (05:45 +0000)]
Change reference from mount(2) to nmount(2), since mount(8) uses
nmount to mount an mqueufs filesystem.

17 years agoFix typo.
Pyun YongHyeon [Tue, 21 Nov 2006 05:41:11 +0000 (05:41 +0000)]
Fix typo.

17 years agoChange reference from mount_devfs(8) to mount(8), and add an example.
Craig Rodrigues [Tue, 21 Nov 2006 04:56:50 +0000 (04:56 +0000)]
Change reference from mount_devfs(8) to mount(8), and add an example.
Remove reference to mount_fdescfs(8), since mount(8) calls nmount()
directly to mount an fdescfs filesystem, instead of forking an
external mount_fdescfs program.

17 years agoAdd TSO support.
Pyun YongHyeon [Tue, 21 Nov 2006 04:40:30 +0000 (04:40 +0000)]
Add TSO support.

Tested by: wilko,  Pieter de Goeje < pieter AT degoeje DOT nl >

17 years agoo Correctly set IFCAP_VLAN_HWCSUM as re(4) can do VLAN tagging/checksum
Pyun YongHyeon [Tue, 21 Nov 2006 04:23:52 +0000 (04:23 +0000)]
o Correctly set IFCAP_VLAN_HWCSUM as re(4) can do VLAN tagging/checksum
  offloading in hardware.
o Correctly set media header length for VLAN.

17 years agoDon't set RL_CFG1_FULLDUPLEX bit. The RL_CFG1_FULLDUPLEX bit in
Pyun YongHyeon [Tue, 21 Nov 2006 04:14:44 +0000 (04:14 +0000)]
Don't set RL_CFG1_FULLDUPLEX bit. The RL_CFG1_FULLDUPLEX bit in
config register 1 is only valid on 8129.

17 years agoUse #ifndef __NO_STRICT_ALIGNMENT rather than
Pyun YongHyeon [Tue, 21 Nov 2006 04:11:31 +0000 (04:11 +0000)]
Use #ifndef __NO_STRICT_ALIGNMENT rather than
#if !defined(__i386__) && !defined(__amd64__) for architectures
with alignment constraints.

17 years agoIncrease WARNS to 3.
Craig Rodrigues [Tue, 21 Nov 2006 02:01:33 +0000 (02:01 +0000)]
Increase WARNS to 3.

17 years agoFix last element of nc_protos[] array to appease GCC.
Craig Rodrigues [Tue, 21 Nov 2006 02:00:48 +0000 (02:00 +0000)]
Fix last element of nc_protos[] array to appease GCC.

17 years agoConvert mount_nfs from old mount(2) API to new nmount(2) API.
Craig Rodrigues [Tue, 21 Nov 2006 01:53:18 +0000 (01:53 +0000)]
Convert mount_nfs from old mount(2) API to new nmount(2) API.

Reviewed by: mohans

17 years agoAdd suspend/resume support. Unlike many other NIC drivers,
John Baldwin [Mon, 20 Nov 2006 23:30:07 +0000 (23:30 +0000)]
Add suspend/resume support.  Unlike many other NIC drivers,
bfe_init_locked() wasn't sufficient to bring the chip back to life, it also
required a call to bfe_chip_reset() during resume.

Tested by: Stepan Zastupov +redchrom at gmail+
MFC after: 1 week

17 years agoDocument that the data field for NOTE_EXIT holds the process exit status.
John Baldwin [Mon, 20 Nov 2006 22:20:04 +0000 (22:20 +0000)]
Document that the data field for NOTE_EXIT holds the process exit status.

Submitted by: Jared Yanovich -phirerunner at comcast.net-
MFC after: 2 weeks

17 years agoSave exit status of an exiting process in kn_data in the knote.
John Baldwin [Mon, 20 Nov 2006 22:17:50 +0000 (22:17 +0000)]
Save exit status of an exiting process in kn_data in the knote.

Submitted by: Jared Yanovich ^phirerunner at comcast.net^
MFC after: 2 weeks

17 years agoFix an integer overflow and allow access to files larger than 4GB on
Lukas Ertl [Mon, 20 Nov 2006 19:28:36 +0000 (19:28 +0000)]
Fix an integer overflow and allow access to files larger than 4GB on
NTFS.

17 years agoOfficially rename archive_{read,write}_open_file() to
Tim Kientzle [Mon, 20 Nov 2006 16:48:04 +0000 (16:48 +0000)]
Officially rename archive_{read,write}_open_file() to
archive_{read,write}_open_filename():
   * Update Makefile to build the files using the new name.
   * Update docs to document the new names, mentioning the
old ones as "deprecated synonyms."
   * The old filenames will be reconnected to the build soon;
I'll soon recyce those files for a slightly different purpose.

17 years agoForced commit to note CVS repo-copy of archive_{read,write}_open_file.c
Tim Kientzle [Mon, 20 Nov 2006 16:45:39 +0000 (16:45 +0000)]
Forced commit to note CVS repo-copy of archive_{read,write}_open_file.c
to archive_{read,write}_open_filename.c.

This is part of a minor file shuffle/name clarification:

   * The archive_{read,write}_open_file() functions are officially
renamed to archive_{read,write}_open_filename(), which more accurately
reflects their function (they open a data stream identified by a filename).

   * Preserve the file naming conventions by renaming the files as
well.

   * Keep the old function names (in the new files) as synonyms for
backwards compatibility.

   * Keep the old files around; they'll be recycled soon for the upcoming
archive_{read,write}_open_FILE() functions (which open a data stream
identified by a FILE * reference).

   * I'm also, of course, making a nod here to MacOS/Windows
case-insensitive filesystems that can't have both _open_file.c and
_open_FILE.c.  ;-)

17 years agoThe clean_map has been made local to vm_init.c long ago.
Ruslan Ermilov [Mon, 20 Nov 2006 16:23:34 +0000 (16:23 +0000)]
The clean_map has been made local to vm_init.c long ago.

17 years agowhitespace fix only
Julian Elischer [Mon, 20 Nov 2006 16:13:02 +0000 (16:13 +0000)]
whitespace fix only

17 years ago- Fix types of "struct vmmeter" members so they are unsigned.
Ruslan Ermilov [Mon, 20 Nov 2006 16:04:41 +0000 (16:04 +0000)]
- Fix types of "struct vmmeter" members so they are unsigned.

- Fix overflow bugs in sysctl(8), systat(1), and vmstat(8)
  when printing values of "struct vmmeter" in kilobytes as
  they don't necessarily fit into 32 bits.  (Fix sysctl(8)
  reporting of a total virtual memory; it's in pages too.)

17 years agoMatch bus space unmap prototype.
Kevin Lo [Mon, 20 Nov 2006 13:21:02 +0000 (13:21 +0000)]
Match bus space unmap prototype.

Approved by: cognet

17 years ago- Make rp(4) 64-bit- and endian-clean as well as work on strict alignment
Marius Strobl [Mon, 20 Nov 2006 12:59:27 +0000 (12:59 +0000)]
- Make rp(4) 64-bit- and endian-clean as well as work on strict alignment
  archs. [1]
- Instead of bus_space_{read,write}*(rman_get_bustag(), rman_get_bushandle())
  use bus_{read,write}*() for efficiency.

Reported by: Peter Losher [1]
Tested on: i386, sparc64
MFC after: 2 weeks

17 years agoMFi386: revisions from 1.641 to 1.643.
Yoshihiro Takahashi [Mon, 20 Nov 2006 12:32:04 +0000 (12:32 +0000)]
MFi386: revisions from 1.641 to 1.643.

17 years agoRemove a redundant pointer-type variable.
Ruslan Ermilov [Mon, 20 Nov 2006 08:33:55 +0000 (08:33 +0000)]
Remove a redundant pointer-type variable.

17 years agoProtect nfsm_srvpathsiz() call with the nfsd_mtx lock.
Pawel Jakub Dawidek [Mon, 20 Nov 2006 07:32:52 +0000 (07:32 +0000)]
Protect nfsm_srvpathsiz() call with the nfsd_mtx lock.

Reviewed by: mohans

17 years agoMFp4: Tweak descriptions in preparation for porting to other members of
Warner Losh [Mon, 20 Nov 2006 06:27:15 +0000 (06:27 +0000)]
MFp4: Tweak descriptions in preparation for porting to other members of
the AT91 arm9 family.

17 years agoUse scheduler API sched_user_prio() to adjust thread's userland priority,
David Xu [Mon, 20 Nov 2006 05:50:59 +0000 (05:50 +0000)]
Use scheduler API sched_user_prio() to adjust thread's userland priority,
use td_base_user_prio to get real userland priority since POSIX priority
mutex may adjust td_user_pri which is an effective priority.

17 years ago1) Fix up locking in nfs_up() and nfs_down.
Mohan Srinivasan [Mon, 20 Nov 2006 04:14:23 +0000 (04:14 +0000)]
1) Fix up locking in nfs_up() and nfs_down.
2) Reduce the acquisitions of the Giant lock in the nfs_socket.c paths significantly.
- We don't need to acquire Giant before tsleeping on lbolt anymore,
  since jhb specialcased lbolt handling in msleep.
- nfs_up() needs to acquire Giant only if printing the "server up"
  message.
- nfs_timer() held Giant for the duration of the NFS timer processing,
  just because the printing of the message in nfs_down() needed it
  (and we acquire other locks in nfs_timer()). The acquisition of
  Giant is moved down into nfs_down() now, reducing the time Giant is
  held in that path.

Reported by: Kris Kennaway

17 years agoWhen counting vm totals, skip unreferenced objects, including
Ruslan Ermilov [Mon, 20 Nov 2006 00:16:00 +0000 (00:16 +0000)]
When counting vm totals, skip unreferenced objects, including
vnodes representing mounted file systems.

Reviewed by: alc
MFC after: 3 days

17 years agoconfig for Gateworks Avila board booting with NFS-mounted root on npe0
Sam Leffler [Sun, 19 Nov 2006 23:58:12 +0000 (23:58 +0000)]
config for Gateworks Avila board booting with NFS-mounted root on npe0

Reviewed by: cognet, imp
MFC after: 1 month

17 years agoadd CPU_XSCALE_IXP425
Sam Leffler [Sun, 19 Nov 2006 23:56:44 +0000 (23:56 +0000)]
add CPU_XSCALE_IXP425

Reviewed by: cognet, imp
MFC after: 1 month

17 years agoGateworks Avila board support:
Sam Leffler [Sun, 19 Nov 2006 23:55:23 +0000 (23:55 +0000)]
Gateworks Avila board support:
o ixp425 support
o NPE network driver (requires Intel microcode)
o h/w qmgr support
o True IDE compact flash over expansion bus
o pci (ath and hifn795x parts tested)
o xscale watchdog timer
o ds1672 RTC on i2c bus
o ad7418 voltage + temp monitoring on i2c bus
o uart

Work done together with cognet, kevlo, and jmg.  Parts of
the ixp425 support obtaine/derived from netbsd.

Reviewed by: cognet, imp
MFC after: 1 month

17 years agochange bus space unmap protoype
Sam Leffler [Sun, 19 Nov 2006 23:47:51 +0000 (23:47 +0000)]
change bus space unmap protoype

Reviewed by: cognet, imp
MFC After: 1 month

17 years agocorrect bus space unmap prototype
Sam Leffler [Sun, 19 Nov 2006 23:46:50 +0000 (23:46 +0000)]
correct bus space unmap prototype

Reviewed by: cognet, imp
MFC after: 1 month

17 years agoelaborate on stepping names; add intel terminology to help
Sam Leffler [Sun, 19 Nov 2006 23:45:33 +0000 (23:45 +0000)]
elaborate on stepping names; add intel terminology to help
people cross-referencing intel docs

Reviewed by: imp, cognet
MFC after: 1 month

17 years agohookup ad7418 and ds1672 drivers
Sam Leffler [Sun, 19 Nov 2006 23:43:29 +0000 (23:43 +0000)]
hookup ad7418 and ds1672 drivers

MFC after: 1 month

17 years agoo define transfer method
Sam Leffler [Sun, 19 Nov 2006 23:42:32 +0000 (23:42 +0000)]
o define transfer method
o attach ds1672 and ad7418, to be cleaned up

MFC after: 1 month

17 years agoi2c and clock driver for Dallas Semiconductor DS1672 RTC
Sam Leffler [Sun, 19 Nov 2006 23:40:58 +0000 (23:40 +0000)]
i2c and clock driver for Dallas Semiconductor DS1672 RTC

MFC after: 1 month

17 years agoi2c driver for Analog Devices AD7418
Sam Leffler [Sun, 19 Nov 2006 23:39:54 +0000 (23:39 +0000)]
i2c driver for Analog Devices AD7418

MFC after: 1 month

17 years agoAdd vm map and object locking to each_writable_segment().
Alan Cox [Sun, 19 Nov 2006 23:38:59 +0000 (23:38 +0000)]
Add vm map and object locking to each_writable_segment().

Noticed by: jhb@
MFC after: 3 weeks

17 years agoPointy hat handed to me by Andrew: had msi_enable on as a default.
Matt Jacob [Sun, 19 Nov 2006 23:24:52 +0000 (23:24 +0000)]
Pointy hat handed to me by Andrew: had msi_enable on as a default.

17 years agoPlay it safe and make MSI and MSI-X an option you have to turn on for MPT.
Matt Jacob [Sun, 19 Nov 2006 23:15:42 +0000 (23:15 +0000)]
Play it safe and make MSI and MSI-X an option you have to turn on for MPT.

17 years agoIf a TMF request fails to start, make sure that we pull it off the
Matt Jacob [Sun, 19 Nov 2006 23:13:12 +0000 (23:13 +0000)]
If a TMF request fails to start, make sure that we pull it off the
pending list and set the state back to free prior to calling mpt_reset
so we don't panic at a later point.

17 years agoDo not call bstp_stop() internally as it clears the running flag which causes
Andrew Thompson [Sun, 19 Nov 2006 23:08:35 +0000 (23:08 +0000)]
Do not call bstp_stop() internally as it clears the running flag which causes
the timer to never be restarted.

Reported by: bz

17 years agoThe global variable avail_end is redundant and only used once. Eliminate
Alan Cox [Sun, 19 Nov 2006 20:54:58 +0000 (20:54 +0000)]
The global variable avail_end is redundant and only used once.  Eliminate
it.  Make avail_start static to the pmap on amd64.  (It no longer exists
on other architectures.)

17 years agoRemove LDR_LOG2_PGSZ, ia64_pgtbl and ia64_pgtblsz. They are part of
Marcel Moolenaar [Sun, 19 Nov 2006 20:04:11 +0000 (20:04 +0000)]
Remove LDR_LOG2_PGSZ, ia64_pgtbl and ia64_pgtblsz. They are part of
a WIP and not used yet.

17 years agoTry harder to not leak src/ build stuff into ports/ environment
Ruslan Ermilov [Sun, 19 Nov 2006 16:28:52 +0000 (16:28 +0000)]
Try harder to not leak src/ build stuff into ports/ environment
by not exposing the MK_* variables that were designed for src/.

Requested by: many

17 years agoBring in a SNMP module to support monitoring if_bridge(4) interfaces
Shteryana Shopova [Sun, 19 Nov 2006 15:42:48 +0000 (15:42 +0000)]
Bring in a SNMP module to support monitoring if_bridge(4) interfaces
via bsnmpd(1). The module implements IETF BRIDGE-MIB as defined in
RFC4188 and a private BEGEMOT-BRIDGE-MIB.

Sponsored by:  Google Summer of Code 2006
Reviewed by:   bz
Approved by:   bz (mentor)

17 years agoAdd .Xr to p_cansee.9, remove .Xr to intro.2.
Ceri Davies [Sun, 19 Nov 2006 13:36:04 +0000 (13:36 +0000)]
Add .Xr to p_cansee.9, remove .Xr to intro.2.

Reviewed by: ru
Approved by: ru

17 years agoAdd manuals for cr_cansee(9) and p_cansee(9).
Ceri Davies [Sun, 19 Nov 2006 13:35:03 +0000 (13:35 +0000)]
Add manuals for cr_cansee(9) and p_cansee(9).
Thanks to Ruslan for the time he spent reviewing these.

Reviewed by: ru
Approved by: ru

17 years agoMove the check for which disc volume we have/want to after we install
Ken Smith [Sun, 19 Nov 2006 08:12:54 +0000 (08:12 +0000)]
Move the check for which disc volume we have/want to after we install
the package dependencies.  Installation of the dependencies may leave
us on the wrong disc volume.

MFC after: 3 days

17 years agoFix typo in comment
Christian S.J. Peron [Sat, 18 Nov 2006 23:38:31 +0000 (23:38 +0000)]
Fix typo in comment

Pointed out by: ru